Tunneling usually causes the dislocation of adjacent faults and further induces disasters such as earthquakes, but the physical mechanism is still not clear. A numerical model of near fault tunnel excavation was established by coupling mechanics of continuous media with mechanics of discrete medium. The discrete element method(DEM)was used to simulate the meso-mechanical behavior of the fault fracture zone, and the finite difference method(FDM)was used to describe the macro-dynamic characteristics of the upper and lower walls of the fault. Based on the multi-factor influence analysis, the physical mechanism of the near-fault dislocation induced by tunnel excavation was explored. The analysis results showed that the fault dislocation caused by tunnel excavation could be divided into four stages: incubation stage, acceleration stage, slowing stage and stability stage. The interaction between the fracture zone and the upper and lower walls was weak, and the rock mass deformation caused by excavation was difficult to propagate through the fracture zone to the other side. The discontinuity and disharmony of the deformation of the upper and lower walls were the main reasons for the fault dislocation. When the tunnel was located in the hanging wall, the fault displacement above the tunnel depth was positive, showing the form of positive fault, while the fault displacement below the tunnel depth was negative, showing the form of reverse fault. When the tunnel was located in the footwall, the result was opposite. The farther the tunnel was from the fault, the smaller the fault displacement caused by tunnel excavation was, and the fault displacement at different depths finally approached to 0. The increase of distance might also lead to the change of fault dislocation form below the tunnel depth. When the tunnel was located in the hanging wall, the fault displacement decreased with the increase of dip angle. When the tunnel was located at the footwall, the fault displacement above the buried depth of the tunnel increased with the increase of dip angle. The fault displacement under the buried depth of the tunnel decreased continuously. The fault displacement at the buried depth of the tunnel first decreased to 0 with the increase of dip angle, and then the dislocation form changed, and the displacement continued to increase.

In this review, the tunnel engineering geological disasters were firstly reported, and then the simulation applications of tunnel engineering related to RFPA were summarized. The following main conclusions were that during the construction of tunnel engineering, the tunnel engineering of main geology disaster types including solid geological disasters, quasi fluid geological disasters and fluid geological disasters; the RFPA numerical methods in tunnel engineering construction related rock mechanics and failure characteristics of acquiring, under the condition of excavation of tunnel damage simulation, bedding rock tunnel excavation simulation, dynamic tunnel under the condition of failure simulation, simulation and partition of deep surrounding rock fracture under the action of seepage tunnel stability analysis were carried out in such aspects as widespread application; at present, RFPA has made important progress in the aspects of calculation accuracy, calculation scale and calculation speed, parallel computation of large-scale solution process, construction of numerical computing cloud platform and so on. It is believed that with the continuous development of technology and program, RFPA numerical calculation method will be more widely used in tunnel engineering simulation.

The application of 4D-LSM in rock blasting was studied. The fundamental principles of 4D-LSM in tunnel surrounding rock blasting were introduced, including system equations, non-reflective boundary conditions and rock fracturing model based on multi-body failure criterion. Aiming at the blasting of tunnel surrounding rock, 4D-LSM was used to establish the surrounding rock vibration analysis model, single hole blasting model and surrounding rock damage model under blasting. On the basis of different tunnel blasting design schemes, these models were processed through geometric simplification and boundary condition simplification to realize the quantitative analysis of the peak load velocity of the key points of the tunnel blasting surrounding rock for the given blasting scheme. Aiming at the problem of single-hole free-surface blasting, the simulation effects of DLSM and 4D-LSM were compared, and the advantages of 4D-LSM in dealing with large dynamic deformation and failure of rocks were demonstrated. The numerical modelling of surrounding rock damage zone depth under different blasting design parameters was realized by 4D-LSM. The safety classification and ranking of different blasting schemes were realized through the prediction and analysis of particle vibration velocity and damage zone depth predicted by 4D-LSM. The optimization of the tunnel rock blasting design could be realized by quantitative sensitivity analysis and ranking of blasting safety based on numerical simulation by using 4D-LSM.

To simulate the effect of grouting to control large lateral deformation and its mechanism analysis, the discrete element pore density flow method was proposed. Through the improvement of the discrete element software MatDEM, the numerical simulation of the fluid-solid coupling process of tunnel grouting was realized. Based on the instance data of the Shanghai Tunnel Project, the discrete element analysis of tunnel grouting under sudden load was carried out. The numerical simulation of the tunnel lateral convergence value was very close to the test result. Numerical analysis showed that as the sudden load increased, the horizontal convergence of the tunnel showed a nonlinear growth trend; the pore density flow method was used to analyze the recovery effect of grouting on the transverse deformation of the tunnel. The results showed that with the sudden load and grouting as the distance increased, the influence of grouting on the recovery percentage of the tunnel lateral convergence decreased nonlinearly. This method can be further applied to the numerical analysis and mechanism research of tunnel grouting under complex conditions.

The large diameter(13.46 m)shield was used to undercross the existing metro tunnel on the North Ring Road-Tianmu Mountain Expressway. This paper started with the long-term monitoring results of the existing metro tunnel, and combined the distribution characteristics of the upper and soft and hard complex strata and the spatial position relationship between the large shield and the subway tunnel, the feasibility of short distance underpass was analyzed, the MJS(Metro Jet System)columns was used to reinforce the surrounding strata of the existing subway tunnel and grouting reinforcement in the tunnel from the ground was proposed. The effectiveness of the overall response measures was verified by the finite difference software FLAC^3D, and the field implementation and the actual monitoring results were analyzed. The practice showed that the scheme of large diameter shield tunneling under the existing subway tunnel with small clear distance in the upper soft and lower hard stratum was feasible, and the measures of MJS ground reinforcement of the surrounding stratum of the existing tunnel and grouting reinforcement in the tunnel were effective, which could provide reference for similar engineering construction.

Aiming at the localized characteristics of engineering rock mass failure, the local updating rule was established for dynamic analysis of engineering rock mass by using cellular automata technique on spatial scale and Newmark integration method on time scale, respectively. Based on a self-developed cellular automata software for engineering rockmass fracturing process(CASRock), the dynamic version, i.e. CASRock.Dyna was developed. Investigating the propagationof elastic waves in the rock mass, the stress wave propagation obtained by CASRock.Dyna was consistent with the analytical results, which verifies the feasibility of CASRock.Dyna for elastic dynamic analysis. The elasto-plastic results obtained by CASRock.Dyna were consistent with the counterparts obtained by the boundary element method, which demonstrated the ability of CASRock.Dyna for non-linear dynamic analysis. Dynamic analysis of rock mass subjected to destress blasting and blasting disturbance was conducted to explore the effect of in situ stress, heterogeneity of rock mass and dynamic parameters on failure degree and scope.

Contact detection and computation efficiency has always been the key problem of three-dimensional discontinuous deformation analysis(3D-PEDDA), which is one of the most powerful numerical methods. Contact detection suffers from low efficiency, contact indeterminacy, and the incapability to deal with concave blocks. The efficiency of the entire DDA program is also one of the most crucial bottlenecks. This paper introduced the multi-cover method and the last entrance plane method to deal with lacking of efficiency and indeterminacy of contact detection, as well as the local convex decomposition method for concave polyhedron. In addition, in order to improve the computational efficiency, 3D explicit DDA was adapted to unify data reading and writing mode, and parallelized by adding precompile instructions. In order to expand to high-performance cloud computing and supercomputers, the new 3D-PEDDA programs were all developed on Linux system, and preliminary experiments were carried out on a simple cloud virtual machine. Numerical examples verified the efficiency and accuracy of the newly developed parallel 3D-PEDDA, and it is promising that it can be easily extended to high-performance virtual machines or supercomputers to realize the analysis of large-scale projects.

A lateral loading model box was used to carry out model tests under two cases: the tunnel was located in loess(case 1)and penetrated by potential sliding surface in loess(case 2). The stress characteristics of tunnel lining in the process of slope deformation were studied, and the laws of slope deformation, pressure change of tunnel surrounding rock and lining moment change were analyzed. Results showed that the surrounding rock near the slope side of the tunnel lining was easy to form a void zone under the two cases. The maximum width of the void zone was about 1 cm in case 1 and close to 2 cm in case 2. The void zone in case 2 was much larger than that in case 1, with a relative increase of 100%; Compared with case 1, the peak value of lining bending moment in case 2 increased significantly, with a relative increase of 159.33%, and the maximum increase was located in the area near the side arch foot of the slope. The slide of the potential sliding surface in the slope significantly increased the possibility of tunnel lining damage.

The stability of geotechnical structures such as slopes and tunnels is often quantified by safety factors. There are three standards for defining factor of safety: the ratio between resistance and driving forces; strength reduction factor and load increasing factor. Aiming at the commonly used strength reduction and load increasing factors, a novel upper bound limit analysis method based on topology optimization theory: discontinuity layout optimization(DLO)was used to study the obtained values considering different conditions. The target functions were proposed considering strength reduction and load increasing. Several numerical examples were tested, indicating the efficiency and reliability of DLO method. Meanwhile, the differences between the factors of safety of strength reduction and load increasing depend on the value itself and friction angle.

To explore the dynamic fracture regularities of coal body under abrupt unloading with consideration of roof subsidence, the discretized virtual internal bond method was used to simulate the dynamic fracture process of the rock-coal-rock body under unloading condition. It was found that the dynamic fracture exhibited three-stage characteristics, namely the initial, the stable and the accelerated failure stage. With consideration of the roof subsidence, a quantitative fracture evolution function was drawn. With the initial in-situ stress increasing, the dynamic fracture in the initial and the accelerated failure stages was more violent, and the duration of the stable failure stage became shorter. The roof subsidence rate had little effect on the initial failure stage. With the roof subsidence rate increasing, the dynamic fracture in the stable and the accelerated failure stages was more violent. With the surrounding rock stiffness increasing, the duration of stable failure became longer. These findings provided valuable reference for the prediction of coal burst.